University of Macau / Faculty of Science and Technology. / Department of Computer and Information Science

Developing a new drug costs up to US$2.6B and 13.5 years. To save money and time, we have developed a toolset for computer-aided drug discovery, and utilized our toolset to discover drugs for the treatment of cancers and influenza. / We first implemented a fast protein-ligand docking tool called idock, and obtained a substantial speedup over a popular counterpart. To facilitate the large-scale use of idock, we designed a heterogeneous web platform called istar, and collected a huge database of more than 23 million small molecules. To elucidate molecular interactions in web, we developed an interactive visualizer called iview. To synthesize novel compounds, we developed a fragment-based drug design tool called iSyn. To improve the predictive accuracy of binding affinity, we exploited the machine learning technique random forest to re-score both crystal and docked poses. To identify structurally similar compounds, we ported the ultrafast shape recognition algorithms to istar. All these tools are free and open source. / We applied our novel toolset to real world drug discovery. We repurposed anti-acne drug adapalene for the treatment of human colon cancer, and identified potential inhibitors of influenza viral proteins. Such new findings could hopefully save human lives. / 開發一種新藥需要多至26億美元和13年半的時間。為節省金錢和時間，我們開發了一套計算機輔助藥物研發工具集，並運用該工具集尋找藥物治療癌症和流感。 / 我們首先實現了一個快速的蛋白與配體對接工具idock，相比一個同類流行軟件獲得了顯著的速度提升。為輔助idock 的大規模使用，我們設計了一個異構網站平台istar，收集了多達兩千三百萬個小分子的大型數據庫。為在網頁展示分子間相互作用，我們開發了一個交互式可視化軟件iview。為生成全新的化合物，我們開發了一個基於分子片段的藥物設計工具iSyn。為改進結合強度預測的精度，我們利用了機器學習技術隨機森林去重新打分晶體及預測構象。為尋找結構相似的化合物，我們移植了超快形狀識別算法至istar。所有這些工俱全是免費和開源。 / 我們應用了此創新工具集至現實世界藥物尋找中。我們發現抗痤瘡藥阿達帕林可用於治療人類結腸癌，亦篩選出流感病毒蛋白的潛在抑制物。這些新發現可望拯救人類生命。 / Li, Hongjian. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2015. / Includes bibliographical references (leaves 340-394). / Abstracts also in Chinese. / Title from PDF title page (viewed on 15, September, 2016). / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only.

published_or_final_version / Chemistry / Doctoral / Doctor of Philosophy

Tse, Ching Man. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2011. / Includes bibliographical references (p. 188-200). / Abstracts in English and Chinese. / Abstract --- p.i / Acknowledgement --- p.vi / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Obtain information on drug targets --- p.3 / Chapter 1.2 --- Drug Design --- p.5 / Chapter 1.3 --- Interface for interaction --- p.9 / Chapter 1.4 --- Summary --- p.10 / Chapter 2 --- Background Study --- p.12 / Chapter 2.1 --- Protein Function Prediction --- p.16 / Chapter 2.2 --- Drug Design --- p.37 / Chapter 2.3 --- Visualisation and Interaction in Biomedic --- p.44 / Chapter 3 --- Overview --- p.48 / Chapter 3.1 --- Protein prediction using secondary structure analysis --- p.52 / Chapter 3.2 --- Knowledge-driven ligand design --- p.55 / Chapter 3.3 --- Interactive interface in virtual reality --- p.57 / Chapter 4 --- Protein Function Prediction --- p.60 / Chapter 4.1 --- Introduction --- p.61 / Chapter 4.1.1 --- Motivation --- p.61 / Chapter 4.1.2 --- Objective --- p.62 / Chapter 4.1.3 --- Overview --- p.63 / Chapter 4.2 --- Methods and Design --- p.66 / Chapter 4.2.1 --- Feature Cell --- p.68 / Chapter 4.2.2 --- Heterogeneous Vector --- p.71 / Chapter 4.2.3 --- Feature Cell Similarity --- p.75 / Chapter 4.2.4 --- Heterogeneous Vector Similarity --- p.79 / Chapter 4.3 --- Experiments --- p.85 / Chapter 4.3.1 --- Data Preparation --- p.85 / Chapter 4.3.2 --- Experimental Methods --- p.87 / Chapter 4.4 --- Results --- p.97 / Chapter 4.4.1 --- Scalability --- p.97 / Chapter 4.4.2 --- Cluster Quality --- p.99 / Chapter 4.4.3 --- Classification Quality --- p.102 / Chapter 4.5 --- Discussion --- p.103 / Chapter 4.6 --- Conclusion --- p.104 / Chapter 5 --- Drug Design --- p.106 / Chapter 5.1 --- Introduction --- p.107 / Chapter 5.1.1 --- Motivation --- p.107 / Chapter 5.1.2 --- Objective --- p.109 / Chapter 5.1.3 --- Overview --- p.109 / Chapter 5.2 --- Methods --- p.111 / Chapter 5.2.1 --- Fragment Joining --- p.115 / Chapter 5.2.2 --- Genetic Operators --- p.116 / Chapter 5.2.3 --- Post-Processing --- p.124 / Chapter 5.3 --- Experiments --- p.128 / Chapter 5.3.1 --- Data Preparation --- p.129 / Chapter 5.3.2 --- Experimental Methods --- p.132 / Chapter 5.4 --- Results --- p.134 / Chapter 5.4.1 --- Binding Pose --- p.134 / Chapter 5.4.2 --- Free Energy and Molecular Weight --- p.137 / Chapter 5.4.3 --- Execution Time --- p.138 / Chapter 5.4.4 --- Handling Phosphorus --- p.138 / Chapter 5.5 --- Discussions --- p.139 / Chapter 5.6 --- Conclusion --- p.140 / Chapter 6 --- Interface in Virtual Reality --- p.142 / Chapter 6.1 --- Introduction --- p.143 / Chapter 6.1.1 --- Motivation --- p.143 / Chapter 6.1.2 --- Objective --- p.145 / Chapter 6.1.3 --- Overview --- p.145 / Chapter 6.2 --- Methods and Design --- p.146 / Chapter 6.2.1 --- Hybrid Drug Synthesis --- p.147 / Chapter 6.2.2 --- Interactive Interface in Virtual Reality --- p.154 / Chapter 6.3 --- Experiments and Results --- p.171 / Chapter 6.3.1 --- Data Preparation --- p.171 / Chapter 6.3.2 --- Experimental Settings --- p.172 / Chapter 6.3.3 --- Results --- p.173 / Chapter 6.4 --- Discussions --- p.176 / Chapter 6.5 --- Conclusions --- p.179 / Chapter 7 --- Conclusion --- p.180 / A Glossary --- p.184 / Bibliography --- p.188